Information Technology Reference
In-Depth Information
Similar to Won and Cremers (2002), the user is
provided with information about how to correct
malformed component compositions, which is not
required with our approach. Our research has been
influenced by advances in groupware technology.
Our current prototypes for semantic portals are
based on Microsoft Sharepoint 2003, a popular
commercial groupware product. The CoCoWare
.Net (Slagter & Hofte, 2002) project is designed to
enable end-user-initiated adaptation of groupware
interactions, based on component compositions.
Such adaptations are accomplished by allowing end
users to compose and extend the behaviour of the
groupware system. To reduce the number of pos-
sible component interactions, the CoCoWare .Net
architecture is limited to four groupware-specific
component types. As in our approach, CoCoWare
.Net users need not deal with low level details of
building connections between components.
sition paradigm to the actual application domain
familiar to the end user. Therefore, it is fair to
assume that the new composition paradigm will
be more usable for our end users. We have success-
fully evaluated this assumption theoretically using
the cognitive dimension framework. An empirical
validation of this assumption that has statistical
relevance will be difficult, though, since we cur-
rently have only limited access to domain experts
(physicians). Our approach has been implemented
in a prototype portal application in the domain of
vision care. We are currently beginning to imple-
ment another portal application in a different health
care domain (palliative care), since we have access
to a larger group of test users. (We have started a
collaborative research partnership with the Victoria
Hospice Society.) Another future research focus
will be on mechanisms supporting the evolution of
portal domain ontologies with respect to existing
portal pages and component libraries.
ConClusion and researCh
direCtions
aCknowledgment
Information delivery and content management
in dynamic, ad-hoc business domains such as
health care have been recognised as a difficult
challenge. The main problem behind the current poor
adoption of EMR software is the high degree of
variability of information content to be managed
on a case-by-case basis. Web portal component
(WPC) technology has great potential to address
this problem, if it manages to provide end users
with an efficient and easy way to compose pages
from interacting WPC components. In this chapter,
we have argued that the user-based component
composition paradigm provided by current Web
portal servers is too limited for applications that
involve frequent changes to portal pages by end
users. In fact, this argument is based on feedback
we received from real domain experts using an early
EMR portal application developed in our lab.
We have developed and presented an approach
to overcoming this problem by shifting the compo-
We would like to thank the anonymous reviewers of
earlier versions of this chapter for their great support
in helping us to improve its evaluation. We thank the
National Science and Engineering Research Council
(NSERC) for funding this research.
reFerenCes
Baron, A. (2003). A developer's introduction to Web
parts. Retrieved August 10, 2007, from http://msdn.
microsoft.com/library/
Buckner, T., Hesmer, S., Fischer, P., & Schuster,
I. (2003). Portlet development guide. IBM. Re-
trieved August 10, 2007, from http://www-128.
ibm.com/developerworks/websphere/zones/por-
tal/portlet/portletdevelopmentguide.html
Cabri, G., Loeonardi, L., & Zambonelli, F. (1998).
Reactive tuple spaces for mobile agent coordina-
tion. LNCS, 1477 , 237-248. Springer.
Search WWH ::




Custom Search